Exploring Gender Bias in Pair Programming Interactions

  • Faculty: Armando Fox, UC Berkeley;  Pablo Fernández & Amador Durán, University of Seville, Spain
  • Students: Nate Weinman (PhD), Aslı Akalin (MS)
Pair programming is not only widely used in industry, but is also a great way to increase student engagement and interest in CS. Therefore, you’d expect that it might be a good way to attract more women to CS, as they are significantly underrepresented. But it’s also well known that both implicit and explicit gender bias—that men outperform women as software engineers—can create an unwelcoming environment for women, thwarting the potential of pair programming to improve gender equity in computing. If a pair of programming partners do not know each other and cannot directly observe each other’s visible gender (for example, because they are pairiung remotely), do we observe evidence of gender bias in the pair’s interactions? If so, how can this bias be mitigated, and how does it affect the work product? Our study aims to answer these questions in a variety of cultural settings through a set of randomized controlled trials with a custom-built software framework that supports pair programming. Here is a short abstract describing the poster displayed below, outlining the project scope and motivation, and extended abstract describing the proposed work we are actively pursuing. We are looking for a student research assistant effective immediately. Initial “trial period” would lead to a paid or for-credit position starting later in the semester or over the summer, and potentially proposing it as an “honors thesis” or even 5th year MS project advised by Prof. Fox and others. For the initial work there are no CS prereqs; as you get farther into the work, fundamental techniques from Data 8 + (CS 88 or CS 61A) will be useful. If you’re interested, read the extended abstract; if you’re still interested, contact Prof. Fox.